Closed immesys closed 7 years ago
After a runtime of ~43 days, BTrDB spontaneously goes out
State changed: CLUSTER STATE e376 LEADER: -- @0 MASH: 0 (0 members, 100.00% unmapped) STATE: Degraded NODENAME STATUS MASH WEIGHT RPREF btrdb-0 down+out 0 100 1.00 ACTIVE MASH PROPOSED MASH [e38c8109] btrdb-0 : 0 - 4294967296 (100) $$ btrdb-0 allcurrent is false $$ btrdb-0 proposedMash is 2, our target mash is 2 $$ btrdb-0 we do not want to advance our mash
This turns out to be a problem with leases and etcd. The solution might be to ensure that BTrDB is connected to all of the etcd EPs. We already added functionality so it will automatically restart if this occurs
After a runtime of ~43 days, BTrDB spontaneously goes out